Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 Geomatics, Landmanagement and Landscape

1 2021

Powiadomienia systemowe

Sesja wygasła!

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: vegetation bias

Sortuj według:

Ogranicz wyniki do:

Two methods to mitigate insar-based dems vegetation impenetrability bias

Tulski Sławomir, Bęcek Kazimierz

Geomatics, Landmanagement and Landscape

2021

no. 2

7--21

Digital elevation models (DEM), including the Shuttle Radar Topography Mission (SRTM), are used in many branches of geoscience as an ultimate dataset representing our planet’s surface, making it possible to investigate processes that are shaping our world. The SRTM model exhibits elevation bias or systematic error over forests and vegetated areas due to the microwaves’ peculiar properties that penetrate the vegetation layer to a certain depth. Numerous investigations identified that the penetration depth depends on the forest density and height. In this contribution, two methods are proposed to remove the impact of the vegetation impenetrability effect. The first method is founded on the multiple regression of two forest characteristics, namely forest height and forest density. The second method uses a lookup table approach. The lookup table and the multiple regression explanatory variables are taken from the freely available datasets, including the forest density data (MODIS_VCF) and global tree height map (GTHM). An important role in this research is played by the Ice, Clouds, and Land Elevation Satellite (ICESat) data. The accuracy tests indicate that the first method eliminates approximately 68% of the elevation bias, while the second method appears to be more effective, leading to almost complete removal of the vegetation bias from the SRTM data. The methods are fine-tuned to the local coniferous forests in Poland. Additional studies are required to finetune the methods for the leaf-off state of deciduous forests. However, a new set of parameters for both methods can be quickly developed for different locations and forest types. Both methods’ functionality and effectiveness can be improved once more accurate forest tree height and vegetation density data become available. These methods are universal in mitigating the vegetation bias from the Synthetic Aperture Radar Interferometry (InSAR) derived model and photogrammetric models.